The 1.59 Å resolution crystal structure of TM0096, a flavin mononucleotide binding protein from <i>Thermotoga maritima</i>

Abstract

The TM0096 gene from Thermotoga maritima is distributed widely among microorganisms, suggesting that it performs a function indispensable for life and/or virulence. Homologs of the TM0096 protein are found in prokaryotes as well as eukaryotes, including the pathogens Yersinia pestis, Listeria, Clostridium, and Bacillus anthracis. TM0096 belongs to the NIFR3 family of proteins, named after the nifR3 gene from Rhodobacter. The biochemical functions of NIFR3 family proteins are unclear, but limited research points to a role in nitrogen metabolism. For example, the photosynthetic purple bacteria Rhodobacter capsulatus experiences an order of magnitude increase in NIFR3 expression under limiting nitrogen conditions.1 The nifR3-ntrB-ntrC operon is translationally controlled by the nitrogen-sensing two-component regulator NTRC. Although T. maritima TM0096 shares 32% sequence identity with the R. capsulatus NIFR3 protein, operon organization is not conserved. A second example is found in Sterkiella histriomuscorum. In a stressful environment (such as limiting nutrient availability) this protist transforms into dormant encysted cells and re-transforms into vegetative cells when favorable conditions return. The dormant cells contain a pool of mRNA transcripts for a NIFR3-like protein.2 This report describes the structure of TM0096 at 1.59 Å resolution. The initial annotation of the TM0096 gene as a TIM-barrel “putative flavin oxidoreductase” is substantiated by the presence of a flavin mononucleotide (FMN) cofactor in the structure. Indications about the protein function gleaned from this model could be helpful in protein engineering applications or the structure-informed design of potential inhibitors. An open-reading frame for TM0096 (Accession code: NP_2279123) was amplified from T. maritima genomic DNA by PCR with a forward primer (5′-GAGGTTAAAGTCGGTCTTGC -3′) and a reverse primer (5′-CCACCTCCTTGATGAAATTATAAAAC -3′) and TOPO isomerase cloned into hexaHis bacterial expression plasmids. Selenomethionine protein was expressed in E. coli BL21[DE3] cells. After lysis, the protein was purified via nickel ion affinity chromatography [50 mM Tris-HCl, pH 7.8, 500 mM NaCl, 10 mM imidazole, 10 mM methionine, 10% glycerol, and 1 mM DTT with a linear gradient from 10–400 mM imidazole]. The protein was then further purified by Superdex 200 gel filtration (Pharmacia) into 10 mM HEPES, 10 mM methionine, and 150 mM NaCl. Crystals were obtained overnight at 21°C via hanging drop vapor diffusion with equal volumes of protein [6 mg/ml, 10 mM HEPES pH 7.5, 150 mM NaCl, 1 mM βME 10 mM methionine, 10% glycerol] and reservoir [15% (w/v) PEG 4000, 300 mM (NH4)2SO4, 28.4 mM β-mercaptoethanol, and 100 mM MES, pH 5.5] solutions. Crystals were transferred to a cryoprotective solution consisting of 20% glycerol and 80% reservoir. Crystals were flash-frozen in liquid nitrogen. Data for SAD phasing were collected at the APS COM-CAT beam line, processed and reduced with MOSFLM, SCALA, and TRUNCATE.4 Selenium atom sites were located with SnB5 and refined with MLPHARE.4 Difference maps were monitored during this process to check and modify the set of Se sites. The electron density map resulting from this phase set was improved by density modification using the program DM4 and used for initial model building with ARP/wARP.6 Refinement was performed with iterative cycles of manual model building with XTALVIEW/XFIT7 and REFMAC.4 Quality assessment of the model was performed with PROCHECK,8 WHATCHECK,9 and SFCHECK10 (Table I). The structure of TM0096 consists of two domains. The N′ domain (residues 5–237) forms a (β/α)8 TIM barrel while the C′ domain (residues 238–309) forms a four helix bundle [Fig. 1(A)]. Gel filtration chromatography and molecule packing within the crystal asymmetric unit suggest that the functional form of TM0096 is a monomer (data not shown). Electron density maps revealed a FMN cofactor and a sulfate ion bound to the C′ terminal face of the protein. The flavin phosphate moiety is anchored by the typical binding motif found in other TIM barrel enzymes.11 The sulfate binding site is likely coincident with the natural ligand binding site of TM0096 given its close proximity to the flavin and interaction with two predicted active site residues (Arg134 and His160, discussed below). Absorbance spectrum peaks at 370 and 450nm confirm the presence of an oxidized FMN cofactor [Fig. 1(B)]. The solvent accessible protein surface in this region contains an electropositive cavity that may permit substrate and/or inhibitor binding (data not shown). A: SPOCK17 ribbon diagram of TM0096 protein with FMN cofactor and three sulfate ions. β-strands and α-helices are shown in cyan and orange, respectively. B: Absorbance spectrum of the oxidized FMN cofactor in the TM0096 protein. inset: Expanded view of same spectrum showing relative absorbance between the protein (A280) and the cofactor (A450). C: Putative catalytic residues and substrate binding site residues are rendered as ball-and-stick figures in yellow and grey, respectively. The FMN moiety is rendered as white ball-and-stick. Figure produced with PyMOL.18 The PredAct program™12 identifies likely active-site residues based on their structural location, polarity and conservation in related sequences. An analysis of TM0096 identifies several residues near the FMN and the sulfate. Whereas residues Cys93, Arg134, His160, and Arg162 almost certainly participate in catalysis, residues Pro94, Lys132, Ile159, Thr163, Gly190, Asp191 probably contribute to substrate binding [Fig. 1(C)]. Consistent with TIM barrel enzymes, these predicted active site residues lie on the C-terminal side of the β-barrel in the loop regions.13 The single ϕ-ψ Ramachandran violation (Ser167) is stabilized by a carbonyl oxygenNH1 hydrogen bond with Arg134. Arg134 is in a critical position to hold the sulfate ion in the putative active site via hydrogen bond interactions with NH2 and NE. A DALI14 search with TM0096 yields significant matches to several other TIM barrels, the best overlay is limited to the TIM barrel of dihydroorotate dehydrogenase (2DOR15) (Z-score = 17.7, 2.9 Å RMSD for 210 equivalent α-carbons). A DALI query with the four-helix bundle yielded weak similarity to the M fragment of alpha-1 catenin (1H6G16)(Z-score = 4.9, 3.3 Å RMSD for 69 equivalent α-carbons). Two observations suggest that the four-helix bundle is not involved in TM0096 catalytic activity. First, there is limited interaction surface between the two domains (808Å2). Second, none of the residues forming inter-domain hydrogen bond contacts represent putative active site residues. However, the conservation among apparent TM0096 orthologs of the C′ domain suggests an important function, perhaps in protein–protein interactions. The structure described here demonstrates that TM0096 almost certainly has oxidoreductase activity. The nature of the substrate is, however, unclear at present. Oxidoreductases play essential roles in a wide variety of cellular functions. The gene content of the T. maritima genome suggests that this free-living heterotroph utilizes a large number of organic substrates.3 Any one of these pathways may require the TM0096 enzyme. Given the broad phylogenetic conservation of TM0096, a role in a central biochemical pathway is likely. Alternatively, since TM0096 is similar to nifR3, it may play a role in nitrogen metabolism. Future experiments will be useful to obtain more information regarding the biological function of the TM0096 protein in bacteria. We thank the staff at the Advanced Photon Source and COM-CAT, for their help during data collection. The atomic coordinates (code: 1VHN) have been deposited in the Protein Data Bank (http://www.rcsb.org/).